diff options
author | Michael Tyler <michael.tyler@arm.com> | 2023-04-12 17:43:17 +0100 |
---|---|---|
committer | michael.tyler <michael.tyler@arm.com> | 2023-06-05 15:57:58 +0000 |
commit | 74921eee924625426429044decefe3673561b174 (patch) | |
tree | 654da1a95e3d42d6af8ad1ff27bb40d77b1fd8c5 /src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst | |
parent | df5d9878008be9b60586df97ebfff197abb5195e (diff) | |
download | ComputeLibrary-74921eee924625426429044decefe3673561b174.tar.gz |
Update CPU kernel implementations and guard directives
Resolves COMPMID-6023
Change-Id: I868975d14c4f98af6716726feda22405a6a4c891
Signed-off-by: Michael Tyler <michael.tyler@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9686
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst')
-rw-r--r-- | src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst/generic.cpp | 16 |
1 files changed, 13 insertions, 3 deletions
diff --git a/src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst/generic.cpp b/src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst/generic.cpp index 2bb22131f7..1f8f863de2 100644 --- a/src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst/generic.cpp +++ b/src/core/NEON/kernels/arm_conv/pooling/kernels/cpp_nhwc_1x1_stride_any_depthfirst/generic.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2020 Arm Limited. + * Copyright (c) 2020, 2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -24,6 +24,10 @@ #include <cstdint> #include <cstring> +#ifdef ARM_COMPUTE_ENABLE_BF16 +#include "bfloat.hpp" +using arm_gemm::bfloat16; +#endif namespace arm_conv { namespace pooling { @@ -41,9 +45,15 @@ void cpp_nhwc_1x1_stride_any_depthfirst_impl( } template void cpp_nhwc_1x1_stride_any_depthfirst_impl(uint64_t, uint64_t, uint64_t, const float *const *, float *); -#if defined(__ARM_FP16_ARGS) + +#ifdef __ARM_FP16_ARGS template void cpp_nhwc_1x1_stride_any_depthfirst_impl(uint64_t, uint64_t, uint64_t, const __fp16 *const *, __fp16 *); -#endif // defined(__ARM_FP16_ARGS) +#endif + +#ifdef ARM_COMPUTE_ENABLE_BF16 +template void cpp_nhwc_1x1_stride_any_depthfirst_impl(uint64_t, uint64_t, uint64_t, const bfloat16 *const *, bfloat16 *); +#endif + template void cpp_nhwc_1x1_stride_any_depthfirst_impl(uint64_t, uint64_t, uint64_t, const int8_t *const *, int8_t *); template void cpp_nhwc_1x1_stride_any_depthfirst_impl(uint64_t, uint64_t, uint64_t, const uint8_t *const *, uint8_t *); |